Live paper test since 9 Sep 2026

SPX Opening Expected Direction Strategy

This strategy reads which way SPX should move in the first minutes after the open at 09:30 ET, waits for the market to prove it, then buys a 0DTE option. One trade a day at most, and plenty of days with none.

Today

Done for today: closed at the target, +$430.

Status as of 26 Oct 2026. Turn on JavaScript for live updates.

How it works

The idea comes from a quirk in how the S&P 500's opening price is made. Everything below follows from it.

1. Read the open. SPX's official 09:30 price is printed before all 500 stocks in it have actually traded, so for a minute or two the index shows a price that's partly yesterday's. SPY, the ETF that tracks it, trades freely before the bell and doesn't have that problem.

   Each morning I compare where SPX printed its open with where SPY says it should be. The two never line up exactly, so I measure today's gap against what's been normal over recent sessions. Only an unusually large gap counts as a read. Most mornings the difference is ordinary and there's no trade at all.

2. Wait for proof. A read is only a forecast. Before buying anything, SPX has to actually move a set distance from its open in the expected direction, and it has to do it quickly. If the move doesn't show up within a short window, the read was wrong or too early, and I let it go.

   This filter costs a little on the days the read is right, but it removes most of the days where it isn't, and those are the expensive ones.

3. Buy cheap or not at all. I buy a same-day SPX option a little out of the money: a call if I expect up, a put if down. Option prices in the first minutes are inflated by the scramble at the open, so I don't pay the asking price. I rest a limit order a little under it and wait.

   If the price never comes down to my limit within a set time, there's no trade, even if the move carries on without me. I never chase. Quotes that are obviously broken, which happen in the first seconds after the bell, are ignored rather than traded.

4. Get out. Every trade has three exits, and whichever comes first wins: a fixed profit target, a wider stop, or a time exit in the afternoon.

   There's no time stop. I tested exits at 15 minutes, 30 minutes, 10:00 and 10:30, and they all made it worse, because many of the winners take an hour or more to reach the target. Cutting trades early shrinks the losers a little but turns those slow winners into losses.

5. The housekeeping. One attempt per day, never more. On half-day sessions the afternoon exit moves earlier to match the early close, and on market holidays it doesn't run at all.

One trade, step by step

How often it trades

Since 9 Sep 2026 there have been 30 sessions I recorded. It traded on 16 of them. On 9 there was no clear read on the open, and on 5 the read never turned into an entry: the market didn't confirm it, or the option never came down to my price. Most of the work is deciding not to trade.

won lost no clear read no entry not recorded

What one trade risks

Everything here is one contract. The most a single trade can lose is what was paid for it, which so far has been at most $2,020. A typical trade was open for about 7 minutes.

Every trade

Times are New York time. One contract per trade, no commissions.

What makes me doubt it

• 16 trades is nowhere near enough. One bad week could flip the total.
• The direction call looks real; the profit is fragile. In my backtest, buying the opposite option at the same moment loses money, so the read on the open is the edge. After reading the market, the key is to also be able to enter the market cheaply and exit timely.
• Paper fills are kinder than real ones. I count a fill when the ask touches my limit. A real resting order might not get filled at that moment.

What would make me trade it with real money

• At least 30 trades, still above zero. Ten trades can be luck.
• A win rate comfortably above break-even. Because the losses are bigger than the wins, it needs to keep winning well over the break-even rate shown above, not just at it.
• Real fills that match the paper ones. A few weeks of one-contract real orders, to see whether my limit orders actually get filled where the paper trades say they would.

When I'll give up on it

If it's below zero after 30 trades, it goes to the graveyard with the others. I also have one change I'm considering, using the options market's own pre-open price to skip bad days. I wrote down the rules for testing it before collecting the data, and I'll decide in January 2027, after about 40 days with a signal.

Changes along the way

• 10 Oct 2026: the end-of-day exit moved earlier in the afternoon. The old exit never actually fired because of a bug. No trade on this page was affected; all of them closed well before then.
• 11 Oct 2026: obviously wrong option quotes around the open are now ignored. The quote feed sometimes shows placeholder prices in the first seconds.
• 14 Oct 2026: on half-day sessions, when the market closes at 13:00, the end-of-day exit now moves earlier to match. Holidays are skipped.

Questions

Paper trades, filled at real market bid/ask quotes, one contract each, no commissions. The trade table is rebuilt every night; the Today box updates live. This site records my own testing and trading. It isn't financial advice. 0DTE options can go to zero within minutes.

Disclosures · Privacy · [email protected]